From: Rex <[EMAIL PROTECTED]> Subject: Re: [time-nuts] Where to get Winchester connections for Efratom FRK? Date: Fri, 23 Nov 2007 00:41:07 -0800 Message-ID: <[EMAIL PROTECTED]>
> swingbyte wrote: > > As an aside, what resolution counter is needed to calibrate the FRK? Hp > > 5328? HP 5370? > > > > > That's kind of an open-ended question. You don't actually need a > counter, you need a source that is known to be at least as accurate as > you want to set the FRK. One question is: do you really have a need for > a level of accuracy that you are trying to work toward? The group, in > general, tries to go towards "because it is there" challenges in > accuracy, stability, and phase noise. It isn't required by everyone. Let me add that you need some sufficient resolution to acheive the intended resolution over reasnoble time. While a GPS receiver may have some jitter, it should be sufficiently stable (unless you made some very bad decissions on antenna position - maximum sky view please). One common method is to measure the time interval between the reference PPS and the rubidium PPS (you can divide it down using TTL divider chain). You log the TI values over time and time-nuts take pride in processing it for stability using Allan deviation plots (which show the low frequency noise process levels) but you may be more interested in the pure frequency error. You should be able to cancel most of the frequency error so that the accumulated phase drift in the TI data for longer runs is almost completely canceled. Large frequency errors is pretty easy, but as you fine-tune it, the noise processes makes it hard to directly compare as you will not be able to directly separate low frequency noise from linear phase drift, i.e. frequency error. This is why it can be handy to actually do the Allan deviation plot. There are many ways to use a GPS and some counter. The principle remains the same, the counter is the tool to analyze the difference between the reference and the device under test. The counter has the resolution, noise process and non-linearities low enought and the reference the stability and precission enought for a particular DUT requirement. Cheers, Magnus _______________________________________________ time-nuts mailing list -- [email protected] To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
